Im still wound up about the drivers license/ identity business, so Im going to rant more and hope it doesnt become preachy. However, if anyone reads along and wants to sing a hymn and take up a collection at the end, well, Im not available for weddings but Im willing to baptize with the grandkids water gun if the collection plate is full.... (Disclaimer: I am not a lawyer. This discussion is more about statistical testing than criminal law.) Consider sitting on the jury of a criminal trial where the avowed standard is innocent unless proven guilty beyond a reasonable doubt. That simple phrase from the American justice system says a lot more than is immediately apparent. First it establishes a boundary, the reasonable doubt, as the dividing line between whether you choose to convict or whether you choose not to convict. In reality, the verdict is guilty or not guilty. Notice that I did not say guilty or innocent. The courts are not to establish innocence, they are to proclaim guilty or not guilty. Someone found not guilty has not been found innocent. Unfortunately, whether you feel that a verdict is just or not may have at much to do with attorney theatrics as it does with a failure to reach a verdict you agree with. The boundary, reasonable doubt, is vague and perhaps cannot be rendered with better precision than whatever floats through the jurors head. In testing a claim based upon carefully gathered evidence, the equivalent statistical test might establish a numerically based standard to decide the claim. For example, Claims might be tested to see whether the evidence is significant. In statistics, significant results means weve reached a conclusion of guilt based on the significant evidence and there is less than a 5% chance that this conclusion is in error. Suppose that the reasonable doubt requirement meant that the conclusion was tested at the 5% level of significance. If we collected similar significant evidence in a 100 such crimes, the verdict of guilty would be correct more than 95 times, and only incorrect less than 5 times. That means as many as five innocent people may be convicted under this standard. We could raise the standard to highly significant with a 1% chance of error. That would imply that with similar evidence to convict, less than 1% convicted would actually be innocent. If you think that innocent people should never be convicted, then you have to wait for Gods judgement. People make mistakes. Evidence is never perfect. Lawyers are not all equal in their presentations, etc. Such is the nature of the system. And remember, reasonable doubt is a vague standard. Convicting an innocent person might be a miscarriage of justice, but a statistician calls it a Type I error. A statistician is interested not in just one case but on what happens on average in hundreds of cases. At first, the obvious answer would be to increase the standard for conviction from beyond a reasonable doubt to Way, way beyond a reasonable doubt, Im sure hes guilty and bet the lives of my family. Statistical analogy: raising the standard from 5% to say 1%. That should at least cut down the number of convictions of innocent people--the Type I error--and it would by as much as 80%. But there is another error that can be made, a Type II error. In law terms, a Type II error occurs when a guilty party (not innocent) is found not guilty. Type I and Type II errors are interconnected. Decreasing the likelihood of a Type I error to cut down on innocent folks found guilty will increase the probability of a Type II error, the error of finding a guilty person not guilty. Conversely, if you are determined that every guilty person be found guilty by decreasing the likelihood of a Type II error you will raise the probability standard that an innocent person may be found guilty. Arguments (or mathematical proofs) connecting this relationship between Type I and Type II errors can be found in many statistics books, but usually require building some background first. Seems like a long way to go for a rant, and some folks dont really care about the criminal justice system unless they are forced into its tentacles. However, if youve read this far, then perhaps you already see how this is connected to my drivers license. Ive seen no significant evidence in the state of Texas that voter fraud is rampant. If its 1 in a half-million votes I would be surprised. Perhaps it would be higher if LBJ were still running for congress. Yet a lot of time and money is being redirected to identifying voters. Stupid voting is a whole other matter that I wont address here although Im inclined to call it a Type III error. We can spend a lot of time, money and aggravation in a futile attempt to insure that no person votes who is not entitled to vote, because unlike the criminal justice system where peoples lives are at stake, voting rules must create perfection, therefore no person not entitled to vote should ever be allowed within fifty feet of the ballot box except to clean up the trash. Such additional requirements formulated to keep un-entitled individuals from voting will inevitably result in the connected error: citizens with every right to vote will be kept from voting. Conversely, making it easier for qualified citizens to exercise their right to vote may result in an increase in voter fraud. I dont know of any state in the country where voter fraud has been identified as a major problem, but I take serious issue with anyone who favors restricting the voting rights of those entitled to vote. Now, if you believe that its alright to restrict voter rights because you are paranoid about fraud, I suppose thats a point of view. But if you think restricting voting rights to accomplish your goal is a good course of action because the people you are likely to push out of participating dont vote the way you think they should, then I dont have the time to waste arguing with you. If politicians in this election year have succeeded in making you think that voter fraud is the most important issue we confront, then theyve distracted you again.
Posted on: Thu, 25 Sep 2014 17:14:51 +0000